Dynamic modeling of the TOR signaling pathway in yeast

This Master thesis presents a complete cycle in the use of models for the study of a complex pathway like the TOR pathway. Previous models were reproduced for the phosphatase part of the pathway, and the assumptions used and the results they provided were studied critically, along with the information content of the experimental data they are based on, thanks to a sensitivity analysis framework. This showed that most parameters can be estimated within high bounds of inaccuracy and that further experiments would be useful, some of which can be directly suggested by the analysis. Literature research gave way to extensions and modications of extisting models in dierent directions of the TOR pathway : Npr1, Sch9, Gln3 and Sfp1 regulations. For Gln3 regulation, new quantitative dynamic data was obtained for its localization upon Rapamycin treatment, and used for parameter estimation. The advantage of this new dataset was assessed. Further study of the models revealed that hypotheses could not be discriminated solely by fitting experimental data, but mutant results also gave important information, which nevertheless need to be taken carefully. A new model for Sfp1 control showed qualitatively correct behavior. Sensitivity analysis on this new model was done as well, which also showed that parameters for the non-phosphatase part of the pathway cannot be accurately estimated with the data available, although the dataset on new interactions appears to provide more information. In conclusion, this work shows the progression through creation of extensions to the TOR pathway, assessment of experimental needs, the obtention of experimental data for parameter estimation, and an assessment of the results of simulations and optimization